High-yield Selection Criteria for Cranial Computed Tomography after Acute Trauma

OBJECTIVE: To develop and prospectively evaluate criteria for selecting head-injured patients requiring cranial CT. METHODS: A two-phased prospective observational study design was implemented at a university ED. Physicians completed a form with 51 variables for 540 patients in Phase I. Ten high-yield variables were identified and prospectively tested on 273 patients in Phase II. Prediction rule performance for identification of patients with abnormal CT scans was determined. RESULTS: The combined criteria had a sensitivity of 97% for CT-scan abnormalities in Phase I; sensitivity was 96% in Phase II. Negative predictive values were 97% and 94%, respectively. Prevalence of disease in Phase I was 17%; prevalence in Phase II was 16%. Had the Phase I criteria been implemented during Phase II, 43 of 273 patients (16%) would not have been scanned, including two patients with positive results who did not require operative intervention. CONCLUSIONS: Fifty-one clinical variables from head-injured patients were narrowed to ten of statistical significance and consistent interpretation for prospective evaluation. Patients with none of these criteria were found to be at low risk of having sustained significant head injury.